Dual identity under threat

Running head: Dual identity under threat

Dual identity under threat: When and how do Turkish and Moroccan minorities engage in politics?

Fenella Fleischmann1, Karen Phalet2 & Marc Swyngedouw2

1 ERCOMER, Utrecht University, the Netherlands

2 University of Leuven, Belgium

Published in Zeitschrift für Psychologie/Journal of Psychology, Special Issue Migration and Integration, edited by G. Horenczyk, I. Jasinskaja-Lahti & D.L. Sam. Volume 221 / Number 4 / 2013 pp. 214-222

Dual identity under threat: When and how do Turkish and Moroccan minorities engage in politics?

Abstract

Drawing on the literatures on dual identity and politicisation, this study relates the political engagement ofEuropean-born Muslim to their dual identification as ethno-religious minorities and as citizens. Minorities’ political engagement may target mainstream society and/or ethno-religious communities. Surveying the Turkish and Moroccan Belgian second generation, our study analyses their support for religious political assertionparticipation in ethno-religious and mainstream organisations, and trust in civic institutions. Its explanatory focus is on the dual ethno-religious and civic identifications of the second generation and on perceived discrimination and perceived incompatibility as threats to theirdual identity. Our findings show that participation in organisations beyond the ethno-religious community is most likely among high civic and low ethnic identifiers, and lower among dual identifiers. Rather than increasing political apathy, perceived discrimination goes alongwith higher levels of participation in both ethno-religious and mainstream organisations.Finally, the perception of Islamic and Western ways of life as incompatible predicts greater support for religious political assertion and lower trust in civic institutions. Implications for the role of dual identity and identity threat in the political integration of ethno-religious minorities are discussed.
Introduction

Due to large-scale immigration in the second half of the 20th century, European societies host growing numbers of citizens who identify with more than one cultural group. Many ethnic minority members develop and maintain dual identities: they are committed to a super-ordinate national or civic identity, which they share with majority members; at the same time, they often maintain a strong sense of belonging to their ethnic and/or religious subgroup (Dovidio, Validzic & Gaertner, 1998; González & Brown, 2006). In their influential paper on politicized collective identities, Simon and Klandermans (2001) argue that dual identification with both the minority group and the wider polity is required for minority members to become politically engaged. Minorities’ political engagement is critical for their societal integration.In many European societies, ethnic and religious tensions focus particularly on the allegedlyproblematic integration of Muslim minorities (Sniderman & Hagendoorn, 2007). Widespread majority views of Muslims as second-class citizens pose a threat to the dual identity of the second generation as Muslim minorities and citizens. Against this background, the present study investigates the hitherto under-researched political consequences of threat to the second-generation’s dual identity. Extendingthe existing literature on dual identity and politicisation, this research has a twofold aim. First, we aimto examinethe role of perceived discrimination and perceived incompatibility as threats to dual identity in minorities’ political engagement. Going beyond levels of identification as precursors of political engagement, our measures of perceived threat addressthe conditions under which dual identities can be developed and maintained.Second, we aim to differentiate between conventional and contentious forms, andbetween mainstream and ethnic or religious forms of political engagement. Thus, we move away from overall levels of politicisation towards different forms of minorities’ engagement in the political domain. Sincedual identifications and identity threat may differentially predict distinct forms of engagement, the broad range of indicators of political engagement in our study provides more nuanced insights into the process of politicisation.

In democratic societies, the political engagement of citizens from allsegments of societyis required for the legitimacy of the political system. Including minorities into this process thus constitutes an important aspect of their societal integration. Moreover, minorities’ political engagements in democratic societies can take different forms, including ethnic and contentious politics. Therefore, the broad range of political outcomes considered in the present study will enhance our understanding of an important, yet under-researched, dimension of the integration of ethnic and religious minorities into European societies.

Dual identification and political engagement

Simon and Klandermans (2001) argue that dual identity is a prerequisite for the politicisation of collective identities. They suggest that only if minorities identify at once with their minority group, which may be defined in ethnic or religious terms, and with the wider society,will they make claims and take action to reduce intergroup inequality. Several empirical studies suggest that dual identification indeed fuels politicisation, over and above other common explanations like grievances (Klandermans, Van der Toorn & Van Stekelenburg, 2008; Simon & Ruhs, 2008; Simon & Grabow, 2010).

While providing some evidence for the importance of dual identity in the politicisation process, these studies have a shortcoming, which is addressed by the present research. Namely, they focus on grievance-based politicisation andthey explain when individual members of low-status groups are willing to take collective action to improve their group’s position (cf. Wright, 2001). While this motivation and the ensuing political action certainly are important from an intergroup perspective, they cover only a small part of the range of political behaviours individuals may engage in. Due to its inherent goal of changing intergroup relations to the advantage of low-status groups, collective actionis likely to meet with considerable resistance from high-status groups. In contrast with such contentious politics,political engagement can also take on more conventional forms, as indicated for instance by citizens’ trust in political institutions. Following a classic distinction between conventional and contentious forms of political engagement in political science (e.g. Tilly & Tarrow, 2006), we study a range of political engagements ranging from most conventional forms (trust in civic institutions),through more activist and therefore potentially more contentious forms (participation in mainstream and ethno-religious organisations), to most contentious forms (assertion of Muslims’ religious rights in the political sphere). Despite these differences, all of the studied outcomes are acceptable or normative types of political action in democratic societies (cf. Kawakami & Dion, 1995) which have been used in the Belgian political context by other groups in the past. In addition to conventional and contentious forms of political engagement, we also distinguish between minorities’ engagement within the ethno-religious community and engagements that target a wider audience across ethno-religious group boundaries.

Dual identity threat: Perceived discrimination and incompatibility

Dual identity, or the integration of ethnic and mainstream belonging, has been related to the psychological and social adaptation of immigrant minorities, and was found to facilitatetheir well-being and to foster harmonious intergroup relations (Berry, Phinney, Sam & Vedder, 2006; Dovidio, Validzic & Gaertner, 1998; Hutnik, 1991; González & Brown, 2006; Hornsey & Hogg, 2000). More recently, however, researchers have asked under what conditions dual identities can be developed and maintained (Fleischmann & Phalet, in preparation; Mahönen, Jasinskaja-Lahti & Liebkind, 2011; Verkuyten &Yildiz, 2007); and they have highlighted possible downsides to dual identity under threat (Baysu, Phalet & Brown, 2012). The latter studies have identified perceived discrimination as a source of identity threat, because the experience of discrimination most clearly signals that one’s minority identity is devalued by the majority (Branscombe & Ellemers, 1998).

As a source of identity threat for dual identifiers, perceived discrimination may have direct implications for their political engagement, which our study will examine further. There is some evidence relating perceived discrimination to political engagement, but the nature of the association depends on the specific political outcome under study. In a previous study among Turkish and Moroccan minorities in four European countries, perceived discrimination was reliably related to increased collective action like demonstrations or petitions, but decreased support for the political assertion of religion (Fleischmann, Phalet & Klein, 2011). Since perceptions of discrimination and unfair treatment by definition imply adversarial attributions as well as illegitimacy appraisals, they are most likely to initiate an emotion-based pathway to collective action (Van Zomeren, Spears, Fischer & Leach, 2004). Thus, they mayfuel contentious forms of political engagement with a view to redressing intergroup inequality.

In contrast, other forms of identity threat, for instance, negative stereotyping, less clearly imply adversarial attributions to a high-status outgroup; and might be differentially associated with political engagement. Concretely, we assessed incompatibility beliefs, opposing Islamic to West-European identities and loyalties,as a distinct source and a most direct measure of dual identity threat. Specifically, we asked to what extent Islamic and European lifestyleswere seen as inherently conflicting by minorities – in line with prevailing majority perceptions of incompatibility (Van Acker & Vanbeselaere, 2011). Since the perception of incompatibility does not imply adversarial attributions, nor does it represent the situation as illegitimate, this form of dual identity threat is less likely toinduce emotion-based political action. Rather, we expect that perceived incompatibility may predict minorities’ disengagementfrommainstream politics. At the same time, perceived incompatibility may underlie some contentious or oppositional forms of political engagement within the ethno-religious community. More specifically, we predict that perceived incompatibility will be most positively related to religious political assertion, and negatively to mainstream forms of political engagement. Perceived discrimination, on the other hand, is expected to go alongwith organisational participation as a contentious, yet effective means of redressing the grievances of ethno-religious minorities.

The present research

Based on the considerations above, we will analyzefour indicators of political engagement: support for religious political assertion, participation in ethno-religious and mainstream organisations and trust in civic institutions.[1] In contrast to previous research on politicisation, which averagedover a range of political attitudes and behaviours (Simon & Ruhs, 2008; Simon & Grabow, 2010), we distinguish between these four indicators. This should provide a more fine-tuned understanding of howminorities’ dual identification affectspolitical engagement, and of the forms of political engagement for which it is most relevant.

As one explanation for different forms of political engagement, we examine the role of dual identity. Dual identity is operationalised as the interaction betweenminorities’ identification with the ethnic and religious subgroups(i.e., ethno-religious identification) and their identification with the city and country of residence (i.e., civic identification). In line with a well-established bidimensional approach of identification and acculturation (Hutnik, 1991), we used separate measures of ethno-religious and civic identities, and added a statistical interaction (after centring) which indicates whether the combination of these identities affectspolitical engagement over and above their additive effects. We consider participants who score high (i.e., above the midpoint of the scale) on both ethno-religious and civic identification as dual identifiers;andwe expect dual identifiers to be most politically engaged (H1; cf. Simon & Klandermans, 2001).

In addition to dual identification, we expect that perceptions of dual identity threat will also predict political engagement. Also in line with Simon and Klandermans (2001), wehypothesize that minorities who perceive their identities to be compatible will be most politically engaged (H2a). Additionally, we expect them to support the goals and participate in the political activities of both their ethno-religious community, and the wider society(H2b). On the other hand, those members of the second generation who perceive their ethno-religious and civic identities to be incompatible and who feel discriminated against should be more likely to disengage from conventional politics. While perceived discrimination will predict higher levels of contentious political engagement within and across ethno-religious group boundaries, perceived incompatibility will ratherpredict political disengagement and/or restrict political engagement to the ethno-religious community (H3). In summary, we do not expect the four forms of political engagement under study to be interchangeable. Instead, we expect different patterns of engagement as a function of the levels of identification with ethno-religious and civic identities, their perceived (in)compatibility and perceived discrimination.

Data & method

Our empirical analyses are based on survey data, collected in 2007 and 2008 among random samples of 18 to 35 year old local-born children of immigrants from Turkey and Morocco in the major Belgian cities Antwerp and Brussels (see Swyngedouw et al., 2008 for a full technical report). Being born and raised in the cities under study, the second generation is most likely to claim dual identities as citizens and members of ethnic and religious minority groups. Turkish and Moroccan immigrants arrived in Belgium from the 1960s onwards mainly as guest workers and due to this disadvantaged social background their childrenencounter severe obstacles in education and on the labour market (Phalet, Deboosere & Bastiaenssen, 2007). The survey, which was conducted through face-to-face interviews, contained questions on integration in all domains of life, including measures of identification and political engagement. We make use of pooled data from the two minority groups (N=1,159).[2]We run separate regressions for thefour dependent variables (OLS for continuous dependent variables, logistic for dichotomous variables).[3]In addition to identification, perceived incompatibility and perceived discrimination as predictors of political engagement, we include the other three dependent variables and important controls. Since we work with cross-sectional data and hence cannot establish the causal ordering of variables, we explicitly aim to reveal patterns of associations between dual identity and different forms of political engagement.

Measures

Four indicators of political engagement are analyzed. (1) Religious political assertionis assessed with fivestatements with which participants could agree or disagree on 5-point scales.[4]The items form a reliable scale (Cronbach’s α = .71) and extend a scale of Muslim politicisation that has been validated in previous cross-national research (Fleischmann, Phalet & Klein, 2011). We averaged the five items and recoded them such that higher levels indicate more support for Muslim political assertion. Like all other continuous variables, it was centred on the midpoint of the scale to facilitate the estimation and interpretation of regression models with interactions. (2)Participation in ethno-religious organizations is measured by a dummy variable, indicating whetherrespondents participated in any (religious or other) organization targeting mainly co-ethnics during the last twelve months (1=yes, 0=no). (3) Similarly, participation in mainstream organizations indicates whether respondents participated in any organization that was not targeting co-ethnics. (4) Trust in civic institutions is a reliable composite scale (Cronbach’s α = .84) averaging oversix items. The items assess trust in the Belgian government, parliament, the king, the municipal administration, the police and trade unions. Participants indicated their trust on 5-point scales, with higher values indicating more trust.

The main explanatory variables are identification, perceived discrimination and perceived incompatibility of Western and Islamic ways of life. We take four types of identification into account: national, city, ethnic and religious. They were measured with the question: “How strongly do you feel you belong to the following groups? To what extent do you feel….?” (i) Belgian, (ii) Antwerpenaar/Bruxellois, (iii) Turkish/Moroccan, (iv) Muslim. Respondents could indicate their degree of identification on a 5-point scale ranging (after recoding) from 1 (very weakly) to 5 (very strongly), with respondents coded as 0 if they indicated that the category of identification did not apply to them. Because nationaland city identification are positively correlated, as are ethnic and religious identification, scores for civic and ethno-religious identification were computed as the mean of the respective two identities.[5]These two types of identity and their interaction are added as predictors, with the interaction indicating whether they affect political engagement additively, or whether political engagement differs depending on specific combinations of ethno-religious and civic identification.

Perceived incompatibility of Islamic and Western life-styles was assessed with three items: “The Muslim lifestyle is incompatible with the lifestyle in Western Europe”, “The values and culture of Western Europe are a threat to Islam” and “If worse comes to worst, the countries of Western Europe will turn against Islam”. Participants indicated their agreement with these statements on 5-point scales. The reliability of the scale is rather low (Cronbach’s α = .53).[6]

Two items measure the frequency of perceiveddiscrimination, asking how often participants were personally treated unfairly or experienced hostility due to their background on 5-point scales (1=never, 5=regularly). Because the two items are highly correlated (r = .57), the average is computed to assess perceived discrimination.

As control variables, we include the ethnic background of the participants, contrasting Turks (1) with Moroccans (0), participants’ gender (1= female) and their level of education (tertiary degree completed or currently pursued =1).[7] Table 1 presents descriptive statistics of all variables included in the analyses; Table 2 shows the bivariate correlations between continuous variables.

- Tables 1 and 2 about here -

Results

In the following, we will present the results of OLS and logistic regressions of religious political assertion, participation in ethno-religious and mainstream organisations, andtrust in Belgian institutions. Detailed results of these regressions (coefficients, standard errors and explanatory power) are presented in the appendix (Table A1). We schematically present the results in Table 3, which shows the significant effects of the main variables of interest (with + for positive and – for negative effects, 0 indicates a non-significant effect).

- Table 3 about here -

With regard to the effects of identification, Table 3 shows that civic identification is positively related to participation in mainstream organisations and trust in civic institutions. In parallel, ethno-religious identification predicts religious assertion and participation in ethno-religious organisations. Importantly, ethno-religious identification turns out to beunrelated tomainstream forms of political engagement, but importantly, is not negatively relatedto them either. There isonly onesignificant interaction between civic and ethno-religious identification, and it relates to participation in mainstream organisations. To illustrate the interaction, Figure 1 shows the predicted probabilities of participating in mainstream organisations for combinations of high and low civic and ethno-religious identification. The results are at odds with our first hypothesis on politicised dual identities: the predicted probability of participation in mainstream organisations is highest for high civic and low ethno-religious identifiers, whereas dual identifiers (high on both) have medium levels of participation, which are only slightly higher than those of low civic and high ethno-religious identifiers. In line with the importance of identification for political participation, those who have low levels of identification with both types of identity are least likely to participate in mainstream organisations. All in all, our first hypothesis that dual identifiers are most likely to be politically engaged (in all domains) finds little empirical support.